Google Translate has been assessed for ‘information retrieval effectiveness’ and was found to be 88% effective for making bilingual Internet searches of non‐medical information (Savoy & Dolamic 2009). Interestingly, while there is a need for quick, cheap and accurate translation of documents into multiple languages, there is limited evaluation of translation software used in health care. Similar to commercial software, these have also been evaluated to give a ranking of quality (Table 1) (Hampshire & Salvia 2010). In competition with commercial translation software, several free online translation portals have become available, e.g. The top commercial translation software for 2012 is presented in Table 1. Commercial translation software is evaluated annually by an independent organization (Top ten reviews 2013). The quality of the translation provided through both sources is variable. There are two sources of machine translators: commercial software and free machine translators available through the Internet. Assessment of quality therefore needs to reflect the context of the source document but also the simplicity of the language. This pragmatic approach may be sufficient for translating non‐essential information but is unsuitable for conveying information in health care, especially when informed consent is required (Hablamos Juntos 2009b). Garcia ( 2010) suggested that rather than assessing quality of translation using criteria for assessing human translation, a more pragmatic approach should be taken so that the translation is understood ‘by the educated bilingual rather than by the professional translator’ (pp. co‐reference structure, semantics of the source language, text style, idiomatic expressions and syntax of the source language and transcription language (Och 2005). To provide quality translation, various factors need to be taken into consideration e.g. The challenge for machine translation developers was identifying large enough corpora for comparison.Īvailability of machine translation does not necessarily equate to higher quality translation. a ‘body of text in the source language paired with its translation in the target language’ (Kirchhoff et al. Machine translators work through referencing the source text to a corpus, i.e. However, it was not until the 1990s that the technology became sophisticated enough for effective translation software to be developed. Since the 1940s, there has been interest in developing automated translation (Kirchhoff et al. However, the results will have translational application to clinical practice where conveying information and gathering experience data is part of standard care. This study focuses on two aspects of research where machine translation may be useful: patient information sheets and data collection through survey methods. One solution to overcome the cost of translating research documents by human translators could be through the use of machine/computerised translation software. It is, however, a challenge to balance fairness and the desire to have a sample representing the population with the limitations of resources to enable comprehensive translation. In the UK, the NHS Research Ethics Committees require justification if only English literate participants are to be included in a study ( ). Verbal information can be conveyed through interpreters, but the process of ‘informed consent’ depends on written information being available, to be used in conjunction with verbal explanations, to facilitate reflection before deciding whether or not to take part (Dixon‐Woods et al. Similarly, for research to be generalizable, it needs to include participants representing the whole population information solely in the native language would not facilitate this. However, to ensure health care is equitable information needs to be available in languages other than just the country's native language. In the UK in 2011, the National Health Service (NHS) spent £23♳ million pounds on translation, which has led some to question whether documents should be available solely in simple English rather than providing translations into multiple languages (Gan 2012). This poses a challenge for public services, because providing information that is accessible to all society therefore requires written documents to be translated in to a wide range of languages and this is costly. This increase can be seen in the change in Census statistics for England and Wales where 87♵% of the population were white British in 2001 (Office for National Statistics 2004), which decreased to 80♵% in 2011 (Office for National Statistics 2012). World migration is higher than ever, not just people voluntarily looking for a better life but also as a result of conflict and environmental disasters (International Organization for Migration 2013).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |